Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -28,6 +28,64 @@ cached_answers = {}
|
|
28 |
cached_questions = []
|
29 |
processing_status = {"is_processing": False, "progress": 0, "total": 0}
|
30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
# --- Web Content Fetcher ---
|
32 |
class WebContentFetcher:
|
33 |
def __init__(self, debug: bool = True):
|
@@ -389,7 +447,7 @@ class AudioTranscriptionTool:
|
|
389 |
# --- Enhanced Intelligent Agent with Direct Attachment Processing ---
|
390 |
class IntelligentAgent:
|
391 |
def __init__(self, debug: bool = True, model_name: str = "meta-llama/Llama-3.1-8B-Instruct"):
|
392 |
-
self.search_tool =
|
393 |
self.client = InferenceClient(model=model_name, provider="sambanova")
|
394 |
self.image_tool = ImageAnalysisTool()
|
395 |
self.audio_tool = AudioTranscriptionTool()
|
|
|
28 |
cached_questions = []
|
29 |
processing_status = {"is_processing": False, "progress": 0, "total": 0}
|
30 |
|
31 |
+
|
32 |
+
# simple search instrad of duck:
|
33 |
+
class SimpleSearchTool:
|
34 |
+
"""
|
35 |
+
Simple search tool that scrapes DuckDuckGo HTML results.
|
36 |
+
Drop-in replacement for DuckDuckGoSearchTool.
|
37 |
+
"""
|
38 |
+
|
39 |
+
def __init__(self):
|
40 |
+
self.session = requests.Session()
|
41 |
+
self.session.headers.update({
|
42 |
+
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
|
43 |
+
})
|
44 |
+
|
45 |
+
def run(self, query: str) -> str:
|
46 |
+
"""Search and return formatted results."""
|
47 |
+
try:
|
48 |
+
# Encode query for URL
|
49 |
+
encoded_query = urllib.parse.quote_plus(query)
|
50 |
+
url = f"https://html.duckduckgo.com/html/?q={encoded_query}"
|
51 |
+
|
52 |
+
response = self.session.get(url, timeout=10)
|
53 |
+
response.raise_for_status()
|
54 |
+
|
55 |
+
soup = BeautifulSoup(response.content, 'html.parser')
|
56 |
+
results = []
|
57 |
+
|
58 |
+
# Find search result containers
|
59 |
+
result_containers = soup.find_all('div', class_='result__body')
|
60 |
+
|
61 |
+
for i, container in enumerate(result_containers[:5], 1):
|
62 |
+
try:
|
63 |
+
# Extract title and URL
|
64 |
+
title_elem = container.find('a', class_='result__a')
|
65 |
+
if not title_elem:
|
66 |
+
continue
|
67 |
+
|
68 |
+
title = title_elem.get_text().strip()
|
69 |
+
url = title_elem.get('href', '')
|
70 |
+
|
71 |
+
# Extract snippet
|
72 |
+
snippet_elem = container.find('a', class_='result__snippet')
|
73 |
+
snippet = snippet_elem.get_text().strip() if snippet_elem else ''
|
74 |
+
|
75 |
+
if title and url:
|
76 |
+
result = f"{i}. {title}\n URL: {url}\n"
|
77 |
+
if snippet:
|
78 |
+
result += f" Snippet: {snippet}\n"
|
79 |
+
results.append(result)
|
80 |
+
|
81 |
+
except Exception:
|
82 |
+
continue
|
83 |
+
|
84 |
+
return "\n".join(results) if results else "No search results found."
|
85 |
+
|
86 |
+
except Exception as e:
|
87 |
+
return f"Search failed: {str(e)}"
|
88 |
+
|
89 |
# --- Web Content Fetcher ---
|
90 |
class WebContentFetcher:
|
91 |
def __init__(self, debug: bool = True):
|
|
|
447 |
# --- Enhanced Intelligent Agent with Direct Attachment Processing ---
|
448 |
class IntelligentAgent:
|
449 |
def __init__(self, debug: bool = True, model_name: str = "meta-llama/Llama-3.1-8B-Instruct"):
|
450 |
+
self.search_tool = SimpleSearchTool()
|
451 |
self.client = InferenceClient(model=model_name, provider="sambanova")
|
452 |
self.image_tool = ImageAnalysisTool()
|
453 |
self.audio_tool = AudioTranscriptionTool()
|